The Global Climatology Analysis Tool
http://www.rap.ucar.edu/projects/gcat
Francois Vandenberghe, Robert Weingruber, Marcel Casado, Shane Swartz, Rong Sheu, Ming Ge, Al Bourgeoi, Terri Betancourt, Scott Swerdlin and Tom Warner
Research Applications Laboratory,
National Center for Atmospheric Research,
PO Box 3000, Boulder CO 80307-3000
303-497-8464, vandenb@ucar.edu
Rick Babarsky
National Ground Intelligence Center,
4353 Seminole Trail, Charlottesville, VA 22911
1. Introduction
High resolution regional climatologies over complex terrain are very valuable and, sometimes, essential for many Defense applications. For those applications, the existing 2.5 degree global model climatology database such as the NCAR/NCEP Reanalysis Projects (NNRP) or the European Re-analysis (ERA 40) are in general too coarse to be very useful, while observational database such as the one maintained by the Air Force Combat Command Center mostly provides surface information with wide gaps in area that are of particular interest for Defense applications, e.g. oceans, Middle-East, Korea, etc. Regional models with proper data assimilation systems offer a capability to combine large scale information from existing global database with local observational information to produce regional circulation flows and environmental thermodynamic properties consistent with local orography. The National Center for Atmospheric Research and the National Ground Intelligence Center have commonly developed the Global Climatology Tool (GCAT), a tool that associates the 5th version of the Penn State/NCAR Mesoscale Model (MM5) with the Four Dimensional Data Assimilation (FDDA) technique to produce fine-scale climatology analyses anywhere in the world.
2. Four Dimensional Data Assimilation
The Four Dimensional Data Assimilation is a Newtonian relaxation-based continuous data assimilation technique, termed "observation-nudging", which allows the sequential insertion of observations into a continuously running mesoscale model with carefully-tuned temporal and spatial weights according to observation times and locations. This data assimilation approach incorporates observations that are available at irregular times and locations, and weights each platform discriminatively. The data that can be routinely incorporated in MM5, for historical analysis as well as real time forecasting, include: conventional twice-daily radio-sounding; hourly surface, ship and buoy observations; and special observations such as NOAA/NESDIS satellite winds derived from cloud, water vapor and IR imageries; aircraft reports of ACAR, AMDAR, TAMDAR and MDCR; NOAA Wind Profilers and NASA/JPL QuikScat sea surface winds.
Figure 1: MM5 RT-FDDA schematic description.
3. Ensemble simulations
In real-time applications, ensemble forecasts are a convenient mean to provide an estimate of the uncertainty associated with that forecast. The same technique can also be applied for analysis purpose. GCAT generates hourly high-resolution (down to 3km) analyses over a region for a time period ranging from an hour to 1 year, (e.g. "March over Iraq"). The mean and standard deviation of the ensemble are computed at a given hour across all the year of the ensemble and plotted to provide hourly climatology means with its year-to-year variability. The runs are conducted in parallel on a state of the art Linux cluster, one year per node, so that large regional climatology can be produced in a reasonable amount of time.
Figure 2: GCAT ensemble concept.
4. Web Interface
The system's setting, monitoring and data display is accessible from any part of the world, with an internet connection through a simple a password protected web interface. Users can, with simple point and click commands, remotely set-up and fire an ensemble simulation on the cluster. The same interface provides a user friendly display capability from which model output graphics can be viewed and output products can be downloaded.
Figure 3: GCAT Web interface: MM5 computational domains location selection.
5. Secondary Coupled Applications:
The high-resolution climatology produced by GCAT can be used to drive secondary application models such as the Second order Closure Integrated PUFF (SCIPUFF) dispersion model, which is part of the Defense Threat Reduction Agency's (DTRA) Hazard Prediction and Assessment Capability (HPAC) toolset. SCIPUFF has been implemented on the cluster. It can run in an ensemble mode or as a single instance, which is driven by the ensemble mean and variance of the meteorological data. Future products may include input files for the HPAC Urban Dispersion Model, the sound propagation model NAPS and the debris trajectory model GTRAJ for which NCAR has already developed interfaces with MM5.
Figure 4: Example of GCAT coupled applications: dual display of GCAT SCIPUFF ensemble output: surface dosage for year 2004 (left) and 2003 (right).
6. Graphics and Output Products:
Classical horizontal and vertical 2-dimensional slice of hourly model outputs for any year, as well as statistical climatology information, such as temporal means and standard deviations computed across all the years, can be displayed and animated. Hourly surface reports, wind profiles and Skewt plots at specified locations (through the point and click functionality) can also viewed for any years and for the ensemble mean. Thumbnail plots are also available for quick simultaneous look at all the years. Several products are currently available for download. These includes formatted input meteorological data files for SCIPUFF plume model, AFCCC wind rose graphic program, Vis5d 3-dimensional display and animation. Some output data can also be downloaded on raster format and be imported in Geographical Information Systems for mission analysis and planning.
Figure 5: Example of GCAT Graphics and Outputs: dual display of surface wind and temperature for year 2004 (left) and the mean of the (ensemble) climatology (right).
Figure 6: Example of GCAT Graphics and Outputs: Thumbnails of plots of wind and temperature at 850hPa for 20 years at a given time.